Automatic Identification of the Rock-Soil Interface and Solution Fissures From Optical Borehole Images Based on Color Features

Abstract

For high-resolution borehole images obtained from digital panoramic borehole cameras, a method for recognizing the rock-soil interface and solution fissures based on color is proposed. Owing to the significantly different colors of geotechnical stratifications, such as the rock-soil interface and solution fissures, compared with bedrock, a self-adaptive geotechnical stratification detection model based on the hue, saturation, value (HSV) color space was established. A binarized image of the geotechnical stratification in the borehole was then obtained using this model. The binary image was filtered to depress the noise, and then segmented. The density of pixels in each segment was calculated to determine the depth, area, and orientation of the rock-soil interface and solution fissures, thereby identifying the geotechnical stratification in the borehole image. Verification of the method with many borehole images and comparisons with corresponding borehole radar images illustrates that it can accurately and quickly identify all rock-soil interfaces and solution fissures in a borehole. The proposed method provides a new automated, efficient, and reliable means of identifying borehole rock structures for engineering applications.